Method and apparatus for detecting obstacle, electronic device and storage medium

Patent metadata

Field	Value
Publication number	US-11468689-B2
Application number	US-202017017508-A
Country	US
Kind code	B2
Filing date	Sep 10, 2020
Priority date	Mar 11, 2020
Publication date	Oct 11, 2022
Grant date	Oct 11, 2022

Abstract

A method and apparatus for detecting an obstacle, an electronic device, and a storage medium. A specific implementation of the method includes: detecting, by a millimeter-wave radar, position points of candidate obstacles in front of a vehicle; detecting, by a camera, a left road boundary line and a right road boundary line of a road on which a vehicle is located; separating the position points of the candidate obstacles according to the left road boundary line and the right road boundary line of the road on which the vehicle is located, and extracting position points between the left road boundary line and the right road boundary line; projecting the position points between the left road boundary line and the right road boundary line onto an image; and detecting, based on projection points of the position points on the image, a target obstacle in front of the vehicle.

First claim

What is claimed is:

1. A method for detecting an obstacle, comprising:
detecting, by a millimeter-wave radar, position points of candidate obstacles in front of a vehicle;
detecting, by a camera, a left road boundary line and a right road boundary line of a road on which a vehicle is located;
separating the position points of the candidate obstacles according to the left road boundary line and the right road boundary line of the road on which the vehicle is located, and extracting position points between the left road boundary line and the right road boundary line;
projecting the position points between the left road boundary line and the right road boundary line onto an image;
acquiring, on the image, regions of interest (ROIs) corresponding to the projection points;
calculating detection results of the ROIs corresponding to the projection points through a depth convolution network, wherein a detection result comprises: a detection result indicting existence of an obstacle or a detection result indicting non-existence of an obstacle;
extracting from the image, based on pre-determined position points falsely-detected by the radar, projection points of the falsely-detected position points on the image;
filtering out, in the detection results of the ROIs corresponding to the projection points outputted by the depth convolution network, the projection points of the falsely-detected position points and detection results of ROIs corresponding to the projection points of the falsely-detected position points;
determining the obstacle in front of the vehicle based on detection results of ROIs corresponding to projection points after the filtration; and
detecting the target obstacle in front of the vehicle according to the detection results of the ROIs corresponding to the projection points.

2. The method according to claim 1 , wherein the calculating the detection results of the ROIs corresponding to the projection points through the depth convolution network comprises:
inputting ROIs corresponding to the projection points into the depth convolution network;
acquiring, in the ROIs corresponding to the projection points, perception results of the position points;
calculating the detection results of the ROIs corresponding to the projection points based on the perception results of the position points; and
outputting, by the depth convolution network, the detection results of the ROIs corresponding to the projection points.

3. The method according to claim 1 , wherein the detecting, based on the projection points of the position points on the image, the target obstacle in front of the vehicle comprises:
detecting a type of the target obstacle according to projection points of the position points on the image; and
calculating a distance between the target obstacle and the vehicle based on the projection points of the position points on the image and predetermined position points occupied by the vehicle.

4. An apparatus for detecting an obstacle, comprising:
at least one processor; and
a memory storing instructions, the instructions, when executed by the at least one processor, cause the at least one processor to perform operations, the operations comprising:
detecting, by a millimeter-wave radar, position points of candidate obstacles in front of a vehicle;
detecting, by a camera, a left road boundary line and a right road boundary line of a road on which a vehicle is located;
separating the position points of the candidate obstacles according to the left road boundary line and the right road boundary line of the road on which the vehicle is located, and extracting position points between the left road boundary line and the right road boundary line; and
projecting the position points between the left road boundary line and the right road boundary line onto an image;
acquiring, on the image, regions of interest (ROIs) corresponding to the projection points;
calculating detection results of the ROIs corresponding to the projection points through a depth convolution network, wherein a detection result comprises: a detection result indicting existence of an obstacle or a detection result indicting non-existence of an obstacle;
extracting from the image, based on pre-determined position points falsely-detected by the radar, projection points of the falsely-detected position points on the image;
filtering out, in the detection results of the ROIs corresponding to the projection points outputted by the depth convolution network, the projection points of the falsely-detected position points and detection results of ROIs corresponding to the projection points of the falsely-detected position points;
determining the obstacle in front of the vehicle based on detection results of ROIs corresponding to projection points after the filtration; and
detecting the target obstacle in front of the vehicle according to the detection results of the ROIs corresponding to the projection points.

5. The apparatus according to claim 4 , wherein the calculating the detection results of the ROIs corresponding to the projection points through the depth convolution network comprises:
inputting ROIs corresponding to the projection points into the depth convolution network;
acquiring, in the ROIs corresponding to the projection points, perception results of the position points;
calculating the detection results of the ROIs corresponding to the projection points based on the perception results of the position points; and
outputting, by the depth convolution network, the detection results of the ROIs corresponding to the projection points.

6. The apparatus according to claim 4 , wherein the detecting, based on the projection points of the position points on the image, the target obstacle in front of the vehicle comprises:
detecting a type of the target obstacle according to projection points of the position points on the image; and
calculating a distance between the target obstacle and the vehicle based on the projection points of the position points on the image and predetermined position points occupied by the vehicle.

7. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions, when executed by a processor, cause the processor to perform operations, the operations comprising:
detecting, by a millimeter-wave radar, position points of candidate obstacles in front of a vehicle;
detecting, by a camera, a left road boundary line and a right road boundary line of a road on which a vehicle is located;
separating the position points of the candidate obstacles according to the left road boundary line and the right road boundary line of the road on which the vehicle is located, and extracting position points between the left road boundary line and the right road boundary line;
projecting the position points between the left road boundary line and the right road boundary line onto an image;
acquiring, on the image, regions of interest (ROIs) corresponding to the projection points;
calculating detection results of the ROIs corresponding to the projection points through a depth convolution network, wherein a detection result comprises: a detection result indicting existence of an obstacle or a detection result indicting non-existence of an obstacle;
extracting from the image, based on pre-determined position points falsely-detected by the radar, projection points of the falsely-detected position points on the image;
filtering out, in the detection results of the ROIs corresponding to the projection points outputted by the depth convolution network, the projection points of the falsely-detected position points and detection results of ROIs corresponding to the projection points of the falsely-detected position points;
det

Assignees

• Apollo Intelligent Driving Technology Beijing Co Ltd

Inventors

• Wang Zhi
• Pan Yong
• Zhao Qinghui

Classifications

• G01S13/867 Primary

  Combination of radar systems with cameras · CPC title

• G06V20/58 Primary

  Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads · CPC title

• G01S13/931

  of land vehicles · CPC title

• B60Y2300/08

  Predicting or avoiding probable or impending collision · CPC title

• G06T2207/30256

  Lane; Road marking · CPC title